 Hello, and welcome. My name is Shannon Kemp, and I'm the Chief Digital Manager of DataVercity. We'd like to thank you for joining today's DataVercity. The future of BI isn't a BI tool, sponsored today by Looker, and brought to you in partnership with Speculous Media. It is a deep dive in continuing conversation from a World Transform podcast, which you can listen to, at the worldtransform.com. Just a couple of points to get us started. Due to the large number of people that attend these sessions, we will be muted during the webinar. If you'd like to chat with us or with each other, we certainly encourage you to do so. Just click the chat icon in the bottom middle of your screen for that feature. And for questions, we will be collecting them via the Q&A section in the bottom right-hand corner of your screen. Or if you'd like to tweet, we encourage you to share highlights or questions via Twitter using hashtag DataVercity. And as always, we will send a follow-up email within two business days, containing links to the slides, the recording of the session, and additional information requested throughout the webinar. Now let me turn the webinar over to Phil Bowarmaster, the host of the World Transform podcast, to get today's webinar started. Hello and welcome. Thank you so much, Shannon. Hi, everybody. I'm very pleased to welcome you to our webinar today. My name is Phil Bowarmaster, and I write and speak about accelerating technologies and the future. I am the co-host of the Future Facing Podcast, the World Transformed, and also our special series, which is called Fast Forward on the World Transformed. On that show, we have conversations with thought leaders who are shaping our future through new ideas and new technologies. Last week, we had on Daniel Mence from Looker, who you'll be hearing from in just a few minutes. And we talked a little bit about this idea that the future of business intelligence might not be what we tend to think it is going to be. Well, speaking as a futurist, I can tell you that people getting the future wrong is not at all uncommon. So I thought it might be fun to start today with a few bad predictions from the past. In other words, how folks have managed to be wrong about the future leading up to where we are now. And believe me, there are plenty to choose from. If you just Google the phrase, bad predictions, you get a bunch of lists. Predictions that were way, way wrong. Predictions that were spectacularly wrong. Many, many predictions that are just plain laughable. But I thought, let's not just look at any predictions. Let's look specifically at predictions about technology. And I found this really fun list over at Forbes. Check out the title, the 15 worst tech predictions of all time. Now, of course, I know that's not what we're here to talk about today. So I'm not going to be able to take you through all 15. But let's just look at a few and kind of get a feel for what we're talking about here. Here's a good one to start with. This telephone got to love the scare quotes on the word telephone. This telephone has too many shortcomings to be seriously considered as a means of communication. And that was from William Orton, the president of Western Union back in 1876, doing a little trash talking on the up-and-coming competition there. Well, telegrams forever, baby, I guess is what William Orton was saying. You know, this telephone thing is never going to catch on. And obviously this is an embarrassingly bad prediction. But just to give Orton a little bit of credit, one trend we've seen in recent years is more and more people do seem to have a problem with the telephone as a means of communication. Texting has become the thing. For some folks, it's all the replace voice communication on the phone. So maybe Orton wasn't completely wrong in every way. Maybe these little mini telegrams that we're sending each other all the time kind of vindicate him to some extent. Here's another one. This is from about 100 years ago talking about this new fangled radio technology. It says, the wireless music box has no imaginable commercial value. Who would pay for a message sent to no one in particular? And we don't know who said this, but we know who they were talking to. They were talking to a fellow named David Sarnoff who was trying to round up investors for his new radio business. And what's interesting to me about this one is I look at this and I said, a message sent to no one in particular being of no value. Well, I said almost exactly the same thing the first time I saw Twitter. And so not only do we get the future wrong, but sometimes bad predictions are cyclical. Sometimes they almost seem to repeat themselves. How about this one? This is a really fun one. And I don't think this one will repeat anytime soon. Nuclear powered vacuum cleaners will probably be a reality within 10 years. And this was said by Alex Lute of 1955. We should note here that Alex Lute was the president of the Lute vacuum cleaner company. So he was a real, say what you want about Alex Lute, a real visionary in his field. Here we are 65 years later, still no nuclear powered vacuum cleaners in sight. So maybe his 10-year window was a little too ambitious. And I think this is a good prediction. This is a good one to keep in mind when you hear things like, when Elon Musk says that we can be back on the moon in five years or we can have a fully functioning Mars colony set up in 20 years. I mean, you know, maybe we can, maybe not, but this is a good one to keep in mind. And I'm certainly not saying we can't do any of those things or that even nuclear powered vacuum cleaners aren't right around the corner. All I'm saying is that ultimately I've got a new item to add to my disappointment list. You know, I used to think that the future was kind of disappointing because we don't have flying cars. Now I think it's disappointing because we don't have flying cars or atomic vacuum cleaners either one. Okay, let's do one more real quick. This is a good one too. This is from Time Magazine in 1966. Remote shopping while entirely feasible will flop. All I can say is I wanted to get a comment on this one from Jeff Bezos but unfortunately he was too busy being the wealthiest human being who's ever lived to return my call so a little disappointed there. I guess we'll just have to let that one stand for now. That one gets to stand. So that takes us now to the topic of the future of business intelligence. And we're going to discover people do get that future wrong but maybe not as creatively as thinking that we'll soon have nuclear powered vacuum cleaners. Actually I think most of the wrong predictions come about more in the vein of the telephone will never amount to anything, radio will never amount to anything, online shopping will never amount to anything. I think what these folks are generally saying is that the future will be like the present. And I think that's probably the same mistake that folks make when they talk about the future of business intelligence. They think the future of BI will be more of the same. And that's the idea that we're actually going to challenge today because we're going to see that the future of BI is not a BI tool. And here to explain that idea to us is Daniel Mintz. Daniel is the chief data evangelist at Looker where he focuses on how people interact with data in their everyday lives and how they can use it to get better at what they do. And today he's going to talk to us a little bit how and why we're getting the future of BI wrong as well as maybe a little bit about what the future holds in store for us. Daniel, welcome. Thank you so much. It's great to be here. So yeah, I'm going to talk about why the future of BI isn't a BI tool and I hope not to get this as wrong as the atomic vacuum cleaner guy. I have a little bit of experience in this and have seen these changes happening in real life. So hopefully they're based on some real observations and I think they will probably resonate with you all as well. So I think we can probably all agree that the world of data is changing. The way that we work with data today, the way that we do data analytics is very different than the way that people dealt with data 10 years ago, 5 years ago, certainly then 25, 30 years ago. And I think there are sort of three big trends that I want to talk through because I think they point us the way to the future. So the first is just that there's been this explosion over the last 10 or so years of applications that businesses use. Now, lots of them are software as a service. They live in the cloud, not all of them, but I think the SaaS apps are probably predominating. So I think that's one big trend that has major implications for how we do data, how we do analytics. The second one is that everybody inside a business wants data. And maybe that was the case 30 years ago. I can't speak for everyone 30 years ago, but they certainly didn't think it was plausible 30 years ago. And today I think most people who work in an information-focused job they expect data, right? It's not just that they want data, but they expect access to data. And then the third, which is kind of a solution to the first two, if you think the first two has sort of problems to be solved, the third one gives us at least a way to solve them is that the infrastructure that we have access to has changed pretty dramatically. So let's take each of those in turn. So the first one, business applications are exploding, right? So I like to just think about how many SaaS applications I've used today. I'm on the West Coast, so it's late morning. I've used an expense tracking app. I've used Google Maps. I've used my email application. I've used my calendaring application and a couple others already, right? And I suspect if you think about it, you've probably done something similar. And if you think about these tools, a lot of them have data in them, right? And so most of the data that we're consuming isn't actually in a BI tool anymore. It's delivered directly into your workflow, right? And this is borne out, or this is sort of proven as well by surveys that industry analysts do, right? So Gartner spoke to people and they realized that the spending that companies will do on traditional BI is really shifting to analytics that are right in people's business operations and digital solutions. And SaaS applications are growing like crazy, right? It used to be that a company might have tens or maybe 100 applications that they used across the business to run the business. Now it's hundreds per department. There are even SaaS applications now that will look at your billing statements to identify all of the SaaS applications that you are using so that you can go, oh, we're actually not using that one anymore. Maybe we should stop paying for it, right? And over the next few years, there will be hundreds of millions of new apps. Most of them won't catch on, but they'll be there. And what that means is that it's worth thinking about the fact that all applications today fundamentally are data applications. There's just data being put into these applications, whether it's an ad tracking platform or Salesforce where you as a salesperson are going in and updating your forecasts. That's data, whether it's something like Zendesk where people are coming to ask for help. Well, there's data in those conversations about how frequently they need help, what kinds of help they need, what the severity of the tickets are, how many tickets you're getting per customer. All of that data is in those applications just automatically. To really drive home the point, I always like to bring up my wife who is a Pilates instructor. And basically all of the time that she spends working where she's not actually teaching is spent in an application doing scheduling, moving students between classes, setting up new classes, realizing that she has an extra slot because somebody canceled. All of that information about her workday is there in that application. So even Pilates instructors are data workers. They're working in data applications. And the problem is with all of these applications, some of these applications like Salesforce maybe is shared across departments, but lots of them are just in one or two departments. And because the data often stays locked away in each of these data applications, there's not a real shared understanding of the business. The way to actually understand your whole business is to draw from all of these different silos to draw all of the data together. So you can imagine, well, if your website goes down, you might find out about that on a log monitoring tool. But the implications for that actually stretch across all of the different parts of your business and the data that you need is in all of those different parts of the business. So if the website goes down, well, you need to turn off your ads because you don't want to be driving people to a website that's down. You might need to identify the high value customers who clicked on an email during that period that the website was down. Well, that information is in your email service provider. You might need to change your pipeline of products because you sold less than you thought you would because of that outage, right? So each of those pieces of information lives in a different application. Usually, those applications are pretty locked away, separate from each other. And the traditional data wrangling approaches that we have that have served us, I'm not sure I'd say they've served us well, but they've served us for a long time, they simply can't keep up with this explosion of SaaS applications, right? We see this all the time where people are trying to manage with thousands of SQL snippets and, you know, thousands of people coming to the data team with requests, huge ETL jobs. The tools that we have just can't keep up. And so what you end up with are these silos, right? Where you're repeating work, you're having to re-clean and reshape the data over and over and over because you have to do it for each of the silos. It becomes really difficult to manage all this stuff. You can't share across departments because stuff is locked away. And of course, it's brittle and unscalable, right? It maybe works as you're a small startup, but as you grow and grow, you pick up more of these applications. There are more silos to deal with. Stuff breaks more often. And, you know, just common sense says that these silos should not be what is driving your decision-making. They shouldn't be constraining your decision-making. If you want to make a decision and the data is there inside your company, you should be able to access it in a way that lets you make good decisions. So that's trend one. Now let's talk about trend two, right? Which is that everybody actually needs all this data that's locked away. And part of this is just our expectations have changed from what they probably were because of all the SAS applications that we use in our regular lives, right, in our non-work lives. So, you know, if you order something online, you get data about what stage it's in and when it's going to be delivered. You know, you pull out the little tool in your pocket and it's full of data, right? About money and ratings on different dishes at your favorite restaurant. We get data about transportation constantly, right? You know, what trains are available or how far away is the Uber that's coming to pick me up or how long is it going to take me to get from here to there because of traffic. Even when we are chilling, we get data, right? We're getting data about what other people are watching, what that's, you know, driving the recommendations that tools give us about what we might want to watch. So that's really changed our expectations. And when you look at who's in the workforce today, what the roles are, a lot of them are data jobs. You might not think of them that way, but they are fundamentally jobs where they need the data in order to do their job, right? It's not just that they're inputting the data, but they really need the data back out in order to do their job well, right? So you might not think of a DevOps engineer as really a data job, but that's the only way you can manage today's cloud servers, right? Just by looking at data about their health, about their utilization, about whether you should reserve some more servers because usage is going up. You know, if you're a product manager, you should be data driven, right? Is the feature that we just shipped, is it being used? Are we delivering on time or are we consistently late? You know, if you're a quantitative marketer, you're probably spending a lot of money on ad campaigns, and it sure would be nice to know from the data which is working and which isn't. So all of these people are hungry for data. They want access to data to do their jobs well. And we know that the companies that use data effectively are far more effective, right? It's not just a sort of side thing. If they use data effectively, it makes the company more successful. Deliverer is a great example for folks who don't know. Deliverer is a huge meal delivery, restaurant delivery service started in the UK and then spread all over Europe. We've been to Europe recently. You've probably had to dodge their bicycle delivery people who are just everywhere. And they are totally data driven. They thrive because everybody in the business has self-service access to the data, where and when they need it. Everything in the business is connected to the data, right? Their teams will reach out to particular new restaurants based on, oh, it seems that we don't have a good curry place in this section of town and people are ordering from farther away and that means our writers have to go further. So let's see if we can find a new place to add to our lists. And what that allows is that their data teams, rather than focusing on constantly answering ad hoc requests, they can focus on strategy, on actually doing deep analysis to help the business grow. And one of the things that's really important to realize if you're a data person like me is that even though data and analytics are really fascinating to me, they're not to most people. Most people don't want data. They want the information that they need to do their jobs better and they want it in a way that is going to make sense to them. That's intuitive. That's delivered to them where and when they need it, right? And that might be a BI dashboard, but it might not. And let's be honest, most BI is still done in Excel, right? And that simply can't keep up with the needs of today's businesses, with the volumes of data that we're dealing with, with the speed and diversity of data that we're dealing with, right? With more data sources than ever before and more people demanding data than ever before. Excel just isn't going to cut it. So the question we should really be asking is, well, how can we create a data platform that serves the whole organization? Right? How can we get everybody on the same platform, a platform that can serve all of the different data experiences that they might need, the diversity of data experiences, and can actually grapple with the huge volumes of data that we're dealing with today? So I'm here to say that there is hope. And the hope comes in part because of the third piece, the third trend, which is that the infrastructure that we have available to us has evolved enormously. So if we look at sort of data engineering trends, big data, the big data revolution, I would say for the most part kind of under delivered. It was a lot of hype. People were, it's going to change everything. I don't know that it changed everything. Maybe you disagree with me, but it didn't change everything for me. But one thing that it absolutely did provide was incredibly cheap, incredibly fast, incredibly powerful databases that usually live in the cloud and are just incredibly easy to deploy, right? We are no longer in the days where you had to buy a database or a data warehouse or a data appliance, as they called them because they were huge, and have it trucked in on an 18-wheeler and put into your server form. That's not generally how we get database access today. Generally, databases are living in the cloud. You just reserve them or just start using them in some cases and pay per use, right? So that's one huge trend that gives us hope. The second is that because of these incredibly powerful databases, the old school way of preparing data for analysis, ETL, Extract, Transform, and Load, is moving more to a world of ELT, Extract, Load, and then Transform, right? And the reason that we did the Transform before the load in the olden days was not because we wanted to, it was because we had to. We had to get the data sort of perfectly manicured and just right before we put it into our data warehouse or our database, because that was the only way that the database could deal with the data. It needed it to be perfectly manicured. It needed it to be a star schema or a snowflake schema. Today, because the databases are not just faster and more powerful and cheaper, but also because they are much more flexible than they've ever been before. They can deal with more kinds of data. We can start to move more towards ELT, right? So we load the data and transform it after it's already in the database, and that's simpler and it's much more flexible. It's much more agile. And the third trend isn't really a trend at all. It's just a fact, which is that SQL, structured query language, the language that analysts have been using to talk to databases for 40 plus years now is the standard. It remains the standard. It probably will always remain the standard or for the foreseeable future at least. So the people who know how to talk to databases that have been doing this for a while are very comfortable with the new tools. Even though they're new tools, they're much more powerful. They still speak the same language. And we see huge growth among cloud databases from a number of vendors, from Amazon, from Google, from Snowflake, from Microsoft, but we just see huge growth across all of these vendors who are providing these incredibly powerful databases and really wonderfully for all of us who are competing with each other to provide the best, fastest, most flexible database. It means that each of them keeps getting better and better. And by 2023, the prediction from Gartner is 75% of all databases will be in the cloud. That might be a private cloud, it might be public cloud, but 75% will be there by 2023. And that gives us the actual scale on the database side to deal with that huge supply of data, of data exhaust that's coming off of all of the applications we're dealing with. It gives us the power and the performance to really do the things we need to do with data to get value from them. It often gives us elasticity where we're only paying for what we use or we can very quickly scale up and scale down. So we're not tied to spending millions of dollars ahead of time in a sort of guessing game of how much capacity we'll need. We're able to get that capacity as we need it and then get rid of it when we don't. It's much easier to use than old school databases. They have a lot more flexibility and they cost really pennies compared to what databases and data warehouses cost in the old days. And I think somewhat paradoxically, or not sure it's really a paradox, but something that people are now coming to realize is that they actually give us more security than databases that are being run inside our own companies. Because the people at Google and Amazon and Microsoft are really good at security. They have security teams of thousands of people. And maybe you happen to work at a company where your security team and your IT team are amazing and they're really great at locking things down. But most companies, that's not the case. And so you're sort of reliant on those people to secure your data if you're the one hosting it. If you've got Amazon or Google hosting your data, they can help you get to the level of security that you need. So we've talked about the three trends, so I guess that brings us to the question if the future of business intelligence is not a BI tool, what is it? And I think it's useful to think about what our guiding principle should be, knowing that there is more data supply than ever before, knowing that there is more data demand than ever before, and knowing that the infrastructure has evolved. What should our guiding principles be? Well, one, we should expect that the growth in SaaS applications will only continue, that there will only be more data exhaust every year, and that our analytics and our analytics strategy has to be able to encompass all of that, has to be able to grow and flex as our needs grow and flex. The second guiding principle is because of that data demand we need to aspire to actually serve the entire data-driven workforce. We don't want to just give them data, nobody actually wants data, they want actionable insight, and that's what we should be providing them, and we should be giving it to them when and where they need it, and that's going to differ from person to person and role to role. And our third is just that we should take full advantage of these modern data warehouses. They're only going to get faster, cheaper, better. We should plan to take full advantage of that because if we don't, there's no way that we can do the first two things. And so our prediction, my prediction is that the future of BI is not only a BI tool, right? I actually think that BI tools are great. They do a few things really, really well, and there are plenty of people who are very comfortable with working with a dashboard or a report from a BI tool, and that's great. We should continue to serve those people that way, but that shouldn't be the end of how we serve people with actionable insights from data. So we should be thinking about how do we provide data to people in a way that's intuitive and purpose-built for them, right? That might look like this for a digital marketing application where this is Looker's digital marketing application where we're pulling together data from Google Ads and Facebook Ads and LinkedIn Ads so that digital marketers don't have to log into each of those tools separately, download a CSV of their campaigns for the last week, stitch it together in Excel, and try to make comparisons across channels. They can do that in one application, but that application looks like the tools that they're already using. It's going to make intuitive, immediate sense to them because it's purpose-built for them. It might look like a Web Analytics tool that mirrors Google Analytics. It might look like an incident reporting tool for DevOps engineers that looks like the tools that they're already using. Or it might look like a custom application for a particular need. We work with a company called Global Payments, which is a huge credit card processor, and they needed to be able to show their merchants which credit card transactions have gone through, which were canceled, which were refunded, and they needed to do that in a way that felt natural and very easy to use for the hundreds of thousands of merchants that they serve. It could not be here as a dashboard and good luck using it. There's no way that that was going to scale to hundreds of thousands of people who aren't data people. They're just people trying to run their businesses. They built a custom application. That is a data application built on top of Looker. We call that powered by Looker. It uses a very sort of purpose-built, straightforward interface that's going to make sense to anybody that isn't going to need a whole lot of explanation, great user experience, and it allowed them to get this rolled out very quickly because by building on a data platform, by not having to build this from the ground up, they're able to take advantage of a lot of the things that underpin today's modern BI tools but present that data to people in a way that makes more sense to them. It sometimes looks like putting data in mobile places. Namely, as a great example, they're a fast-growing HR information system company. They were looking sort of up market and seeing that the biggest players in the space, the work days, and the ADPs have really top-notch analytics available to HR professionals. Well, they needed that if they were going to compete up market and so they came to us at Looker and chose to build their HR analytics with Powered by Looker. For a lot of HR professionals, that might mean looking at that on their phone, and that's fine. A true data platform can provide that data there because that's where it's needed. Sometimes it doesn't even look like a separate application. The data might just come find you in the application that you're already using. Looker's Slack application does exactly that. It leverages the same tools, the same underpinnings, the same foundation in the data platform, but it delivers the data experience right in Slack. I want to be able to ask questions of my data without leaving the tools I'm already using. A lot of the time that's the right solution. That's actually the way to meet the data demand. When we ask CIOs, what part of your technology stack do you think is most important to help your business differentiate and actually win, well, they talk about data. We really think that the future of BI isn't just a BI tool. It's a data platform because a data platform can provide the unified and trusted data that we need to bring together all of that disparate data from all those different sources, make sense of it, turn it into something useful, and turn it into something trusted. With data trust is a really big deal. The only thing worse than having no data is having wrong data. So it's really important that the data that we're putting in people's hands is right. A data platform has an architecture that takes full advantage of the amazing ecosystem that surrounds us and can sort of flex and change as that ecosystem evolves so that we're constantly able to take advantage of the newest best tools. And a data platform can support all of the different kinds of data experiences that we want to provide in all of the different media so that we can meet people where they're at and get those actionable insights into their hands. This is a little bit about sort of how Looker's data platform looks, right? You can have sort of any SQL database underneath. You pull you together your disparate data sources into that SQL database or into a few SQL databases. Put Looker on top. And Looker provides that platform for data which can support business intelligence. It can support embedding and customization into other tools. And it can support the business operations and workflows that are needed to, you know, where people are already working and they want to bring data in. And so that gives us those robust analytics. It gives us governance and unified metrics where we're not, when I ask for lifetime customer value and I want to see lifetime customer value by acquisition channel I don't want to have to know what the current definition of lifetime customer value is. That's not something that I as a person should have to track. Somebody else should be able to take that piece of information, take the agreement across the company about how we're defining lifetime customer value and put it in the data platform, put it in software because software is really good at keeping track of those things. And then when I go and ask for that piece of information the platform can give it to me and it can give me the right one that everyone's agreed on. That's a really critical change in how we do data, right? If you're in Excel you get to make up your own formulas. The data is right there and you just make up your own formulas. That's pretty dangerous because when you get in a meeting with somebody else who has their own Excel workbook or their own BI workbook and they have also done that and then a third person comes in with their own BI workbook the meeting stops being about how to make business strategy and it starts being about how to untangle the mess of formulas that we're using and why our numbers differ, right? So having a centralized data platform that supports all of our data usage is really critical to keeping everybody in the business on the same page. And so the way that we do this, I'll just do a little plug for a looker because the way that we do this I think is pretty unique and I can say that as somebody who is a customer of lookers for three and a half years the thing that really grabbed me immediately was looker's modeling layer. This is something that I think kind of went out of fashion with self-serve analytics but a modeling layer, a semantic layer, whatever you want to call it is absolutely critical. That's the thing that really codifies our business metrics. When we come together and say this is the actual definition of what a customer is, what lifetime customer value is, what our sales regions are, those things don't live in the data, those are business decisions that we make as a whole company and once we've agreed on them the ability to write them down in one centralized place in one flexible modeling layer where they're not locked away forever, where they can be examined and audited and we can make sure that they're still current and then we can change them when we need to. That's really critical and that's the way that looker comes out this problem. We do it using modern tools like Git so that if something changes, if the business logic changes that's great, you update it but maybe you want to see when it changed and who changed it and why they changed it or maybe even roll back the change. With version control you can do that. The second piece is that looker leaves your data in your database. Part of the promise of these amazing, fast, powerful, cheap databases is that they can do a lot more of the work. How would you extract the data from those databases to put them into a proprietary data engine? That's something that we had to do in the old days, not something we have to do today and so by leaving your data in your database we can leverage all of the power that you've invested in and make sure that everybody has access to all of the data that they should be allowed to all the way down to the row level detail which is often really critical to understanding the data. The third piece is our web architecture. Looker was born on the web and that means that we're built to take full advantage of modern technologies like RESTful APIs. Looker has what we call 110% API coverage meaning that you can do everything that you can do through the interface via our API but you can do even more via the API because there's something that you just wouldn't want to do in the interface but are really easy and useful to do via the API. The second piece is that sharing in Looker is as simple as copying and pasting the URL. If I take my URL and I'm doing some analysis and I find something interesting I can send that to Phil and he'll pop it into his browser, open it up and take a look and he's right there. He's right where I was and we can continue the discussion. And the companies that are finding value from this it's not a single point solution that serves one department or one kind of business in the gamut from brand new fast moving companies like Deliveroo and Casper and Buzzfeed to venerable old brands like Hearst and Cisco and Fox. These are companies that have been around for a long time but they're embracing this new way of doing data and seeing great results from it. Just to close, now that I've told you what the future of BI is, that it's not only a BI tool I just want to say if you're on the West Coast or if you're really interested in this please come join us at Join our multi-day user conference in San Francisco this year. November 5th through 7th it's going to be a great time. There's going to be a ton of learning more than a thousand folks will be there and it's just the best way to find out all the amazing things that people are doing with data so that you can do those things too. And I think now we're going to go over the questions. If you want to find more info you can definitely do that on our blog or on our website. You can reach out to us at hello at looker.com but I think now we're going to go to questions. All right. Actually we've got some questions already coming in so why don't we just jump into those, Daniel? The first one comes to us from Alexander and this one popped up around the time you were talking about the data hungry workforce and the question I'm not sure if it was rhetorical or not but I'll give you a chance to respond to it. The question is death comes to the enterprise data warehouse so is that what we're seeing? Is it the death of the data warehouse? I think I would say evolution, rebirth, growth as a new thing. No I think enterprise data warehouses are changing and flexing I think is the idea of an enterprise data warehouse that's trucked into your server form that's probably going away but it's evolving into a cloud data warehouse or cloud data warehouses. One of the other things that's happened just because technology advances is not only have storage and computing costs gone down enormously which makes these very fast cheap databases possible but network costs and network have gone down and network speeds have gone up and so while data is heavy and it's still better to have data close to where it's going to be processed it's a lot easier to move big chunks of data than it ever was before so I don't think the enterprise data warehouse is dead I think it's just changing and evolving with the times. Yeah I think you could list depth of the data warehouse as one of those bad predictions that keeps coming up. Yeah for sure. Oh yeah it's just going to be a data lake which turns it into a data swamp. I always joke that a data lake people are like what exactly is a data lake? I say oh well a data lake is just IT's way of making you go away. You come to that and you say hey I have some data where should I put it? And they say oh just put it in the data lake and you say really will I be able to get it back? Oh yeah yeah sure yeah. Do I need to write up some structure and say no no no just throw it in the data lake. It's because they know 95% of the time the stuff that you're going to throw in the data lake is never going to get used and they'd like you to go away and you know I think data lakes I think we actually are seeing some interesting trends there. There are tools like Athena from Amazon which the biggest data lake of them all is probably S3 Amazon's hard drive in the sky right and Athena is this really great tool that actually lets you start querying the data right in S3 you don't have to move it out. And what we tell our customers is that Athena is an amazing tool for huge volumes of data without moving them to understand what in there has value. Now would I want that to be my main querying tool and keep my data in S3? Probably not. You know I think it's not as performant and not as cheap as a more dedicated data querying tool but it's an amazing way to figure out what actually has value so you can then move that into a data warehouse. Absolutely. I just like to remind everyone listening if you've got a question for Daniel please enter it in the Q&A box. I see there's some questions going on in the chat but we don't know if those are intended for us to answer or not so if you've got a question for us please put it up there in the Q&A box on the screen there. I've got a question related to the data hungry workforce for you Daniel and you know we talk about the fact that everybody needs data. So you've got this data hungry workforce. You've also got obviously data hungry customers and let's talk a little bit about the relationship between those two things because at Looker you've got to be looking at both of those things it seems to me it's both the needs of your customer right who has this data-centric workforce and then of their customers and that's what's driving the need for that to begin with. So how do you manage that handoff between those? That's a great question. So the term that folks tend to use to describe that is embedded analytics to take data that they control and turns it around and shows it to their customers or their suppliers or their partners to help them do better and so Looker did not really intend to be in an embedded analytics solution. It wasn't our main focus but what we realized was that customers were doing exactly that with the platform. They realized before we did that the data platform could work just as well as an embedded analytics solution as it did for internal analytics it works just as well for presenting data to your customers as it does to your colleagues and so embedded analytics our solution is called powered by Looker is actually a huge part of our business it's a really critical and growing part of our business where people are realizing that they need modern tools in order to embed analytics right into their platforms whether that's you know HR company like Namely who's showing HR professionals the turn rate and the ramp time for the employees that they're managing whether it's folks like HP printing they do printing for very high and needs like like wine labels and they are fighting counterfitting of those and so they use data to provide they provide data back to their customers so that those customers can track down counterfeits you know whether it's you know POS systems, point of sale systems that serve merchants or like global payments the credit card processor I showed right all of these companies are sometimes they're both and they're providing data to their colleagues as well but the main thing that they're doing is providing data to their customers and partners because that helps them get more value and it often creates a whole new revenue stream for the company to be able to maybe sell and upsell you know an upgraded version of those analytics to their customers right so we got a question from Elizabeth here she says this is the first time I'm seeing this product is this a data virtualization kind of solution it's a great question so this is another thing I wouldn't describe looker as a data virtualization product but customers often use us that way in the sense that the the sort of data that lives in multiple places is all going to be sort of materialized inside looker right you can have a dashboard in looker that's powered by ten different databases where each tile is hitting a different database to pull that data together into one place and because our semantic layer or modeling layer is very lightweight it's not data itself it's just the formulas that make sense of that data that can be moved around and sort of live very easily in the cloud on a very lightweight server it's the databases that are doing the work and you know there is some capability with more coming to actually do some of the sort of key things that a virtualization layer can do in terms of pulling data together and you know maybe you don't just want one tile from one database and another from another database but you actually want to join that data together and make it you know visible or maybe you want to say oh I want to see you know daily ad spend and I want to see daily website traffic but those things live in different databases and I want to put those on the same chart that kind of thing you can already do in looker and there's more coming in that in that vein. Okay just want to remind everyone if you've got a question for Daniel just pop it into the Q&A box there I noticed in the one of the looker customer stories I read it might have been the global payments one that you talked about there was an example of how they went from more than 80 reports to fewer than 10 that could do the same thing so you go from you see these you know kind of heroic systems that companies put in place to try to solve these problems and something much simpler can be put in place to get 10 reports that did the same thing probably actually did more than that I couldn't get over that stat that you showed from Forrester talking about how 60% of businesses are still using Excel for their BI I just kind of cringe when I read that it's just kind of painful to think about and I'm wondering how much of the message that looker brings to businesses starts with kind of a hey folks look it does not have to be this hard right is that the message yeah I mean the way I put it internally to our sales team is that a lot of the time what you're doing is you're telling people how they can clean up the mess right you you get there and they are in pain because they have a mess and I think that is a core part of our message I mean that's actually something that when I talk to our professional services teams I hear about all the time which is you know they say hey we bring out a new customer and that customer is using you know some existing BI tool Tableau or click or something like that and they say hey we have thousands of workbooks floating around the company how long will it take how many hours will we need from you to convert all of those workbooks into all those reports and the professional services team gets to deliver the happy news that thousands of workbooks does not translate into thousands of reports right the reason that all of those workbooks exist is not because each of them is providing value it's because they're duplicative because people can't find the thing that they need and so they recreate it many of them when you actually look are you know disastrously out of date and sometimes still being used and and so that data sprawl that happens sort of inevitably when you've got a sort of workbook data economy is is a problem and something that that look is very good at cleaning up because by centralizing those definitions and making them reusable you can really cut down on the amount of duplication you can really cut down on the amount of sort of disagreement between different data sources or data workbooks which you know having thousands of workbooks is bad in and of itself but it's really bad when you have different people in the business relying on different workbooks that say different things about the same data that's where things really get bad right I like to say that when when technology fails you tend to know it's pretty noisy about it right if your website goes down someone's going to tell you pretty quickly data tends to fail silently which is way scarier because you know you think you know something you you know analyze the data and you see the trend and you go great let's plan on this and then six months later you're saying man things are not going that well go back and look at that analysis and you realize that something was done wrong in that analysis and so the data failed you you thought you were making business strategy based on good data but in fact you were making it based on wrong data and you know maybe it takes six months to figure that out maybe you never figure it out that that was the cause but the data has failed you right okay so we've got a question from Pam and she's looking back at slide 42 and asks but won't data standards be essential for the 100% in database part of the platform to leverage data across database silos? I'm going to go back to slide 42 so yeah I mean absolutely the data definitions in Looker are centralized and reusable there I mean I think the problem with using things like sequel or workbooks is that the in both cases the transformation that's happening the way that the data is being shaped is kind of impenetrable it's not easy to audit or understand right in sequel you probably end up with a 200 line sequel statement to pull the data out and you know if you come back to that a week or two later you go what was I doing? I'll just rewrite it because it's easier to do that than to try to understand even what you were doing let alone sequel that somebody else wrote with a workbook you know you can do a ton of transformation and then to be able to back that out and see well why someone changed this value because they were quote unquote cleaning up the data but how why why did they do that? And so Looker's approach is to make those transformations to make that pathway from raw data to useful actionable insight very transparent to make it reusable to make it version control to make it centralized and to make it modular fundamentally so that you define lifetime customer value once and then when you want average lifetime customer value you don't have to redefine lifetime customer value you just refer to the existing definition and then when you want you know cumulative lifetime value you can do that too right and so you're just reusing the same chunks of logic without redefining them and that harks back to a software engineering term called dry right don't repeat yourself and the reason is a very good one which is when you repeat yourself you make mistakes and you change things and so by writing something once and then reusing it many times you make sure that all those definitions are the same so that's why Looker takes that approach Okay we've got a question from Angela she says it is an integrated solution that does ETL, ELT OLAP and visualization I know this is a question to you but off the bat I'm going to say ETL and ELT take on a different meaning when you're leaving the data in the database I would assume Yeah I think that's right I mean I think let's think about how to answer that so Looker absolutely does a ton of data transformation it is sort of ELT transformation because we're doing it in the database and I don't mean to suggest that companies that are using Looker and have complex data pipelines don't have other pieces in their sort of ELT pipeline and in fact the way that I describe things is I say you know we came from a place of pure ETL we're not in a place of pure ELT yet what we're in instead is sort of like E little T L T little T it's this ongoing process of constantly continuing to transform the data as you get closer and closer to run time to query time and so Looker tends to handle or customers tend to use Looker for the last pieces of that transformation the things that they want the most flexibility with the most insight and then they'll use heavy weight ETL tools before that and it's a very normal process and one that I don't think is going away but yes Looker does do a ton of transformation it gives you all of the sort of transformation power of SQL but in some ways actually it gives you more transformation power than SQL because you can do things in Looker and Look ML which is our modeling language that you'd never want to do in SQL even though Look ML is built on SQL it just sort of a a markup language on top of SQL so it's very easy for people who already speak SQL to pick up since Looker ends up writing the SQL for you there's just stuff that you there's SQL that Looker can write that does amazing things to data that you simply would never want to write yourself I mean even things like pivoting data you know to make a pivot table if you've ever tried to do that with pure SQL good luck it's pretty gnarly but Looker can do that with the click of a button right so there's things like that where Looker is doing even more transformation than SQL is capable of in terms of OLAP we're not doing sort of traditional OLAP and I think part of that is because a lot of modern databases don't need that in the same way but we are you can sort of build your cubes with Looker just by pointing and clicking and there is a fair amount of functionality as well for sort of pre-building aggregates and things like that and so there's things sort of more modern versions of OLAP I would say but I don't mean to suggest that Looker does everything it's not the slices-dices tool that they used to advertise on late night TV but it does a lot and I think the key thing that it really does is it's that central piece in the data sort of ecosystem that can bring all of the data sources together and then support all of the data experiences whatever those might be and so that piece that as long as all of your data is flowing through Looker what you get is the sort of correct business metrics you don't have to worry about is this data right? Okay now we've got a follow up question from Pam which she put up while you were talking so you may have just answered this but I'll give you a crack at it anyway just in case there's more you want to say on this Pam asks so we'd not change things at source, re-common types of data across silos rather fix it with the ELT approach in Looker? Yeah I think that's exactly right is that you can do a lot of the cleanup in Looker and in database stuff that you used to have to do before the data got into the database because it had to be in that perfect star schema to give it an easy example you know BigQuery, Google's data warehouse can handle and actually likes deeply nested data which is not something that most data warehouses like they would force you to flatten that out but BigQuery is perfectly happy with nested data and can query it in place and Looker can write those unnest queries for you at runtime and they're incredibly performant. Another example is Snowflake has really powerful JSON parsing so you can load JSON into a structured database into Snowflake and then query it right in place so you're not having to do all of that prep work to flatten out that JSON or flatten out that nesting beforehand. So yeah I think there's just a lot more capability in the databases and because Looker is built to sort of fully leverage the native capabilities of each database and Looker writes different SQL depending on which database it's connected to or databases Looker speaks about 45 different dialects of SQL fluently it really can take full advantage of all of that power. Absolutely. Okay so let's wrap it up with one last question about the future. Now when you got going you talked about three big trends that you're tracking. One is the software as a service apps tracking. Second was everybody needs data and the third one was that the infrastructure has evolved. Now if we want to look ahead to the future can you identify trends that are currently emerging or that may once again put us in a position of needing to re-evaluate where things go next or they're up and coming trends in other words and in light of those trends do you think the future of BI will continue to be a data platform that's a great question. I think the sort of thing about a data platform that is not trying to be a single vertically integrated tool that does it all but rather that is built to take full advantage of the ecosystem is that it can grow and flex as the ecosystem changes. I don't know what the state of the art data processing tool will be 5 years from now 10 years from now but it will be faster, I know will be cheaper, I know will be more flexible than the ones that we have now and if you're locked into a tool that comes with its own data processing you're basically betting on that one company to come up with the next amazing thing. Whereas when you have a data platform you're betting that there will be an amazing thing in the ecosystem and that you should be able to take advantage of it and so you know I think that sort of flexibility is why I would bet on a data platform over a BI tool because it can grow and flex with the environment as it changes. In the end maybe the data platform is a more future ready approach than what we had before. Yep. Alright, Daniel, great talking with you today. Thank you all for your participation and at this point I think we'll hand it back over to Shannon. Phil and Daniel thank you so much for this great conversation and presentations today and just to answer the most commonly asked questions here just a reminder I will send a follow-up email by end of the day Friday to everybody with links to the slides and links to the recorder from today recording from today along with the information from Phil and Daniel. Thank you both so much. Thanks all of our attendees for being so engaged in everything we do. Love the chat love the questions that came in. It's just so fantastic. Hope you all have a great day. Thanks again. Thanks, Brian. Thanks, Daniel. Thanks Phil. Thank you, Shannon.